A study on rainfall index suitable for sediment disaster risk assessment by geological features

Yasuhiro NOMURA, Soichi KAIHARA, Hiroaki TAKEMOTO, Taro UCHIDA and Hiroaki NAKAYA

Abstract

Early-warning systems for natural disasters are important tools for disaster risk reduction and for achieving sustainable development and livelihoods. Principal methodology of early warning systems is to set a criterion for occurrences of debris flows or multiple slope failures based on several rainfall indices (60-min cumulative rainfall and soil-water index (SWI)) in all 5 km grid meshes over Japan. Parameters of the tank model for SWI are unified as those of granite around the nation. In this paper, deferent parameters proposed by Uchida et al (2016) based on the area’s geological features are tested for SWI. Modification accuracy of early-warning is examined in model areas. The result shows that the SWI values with tested parameters are smaller than the original ones by approximately 20-30%. Although the relative relationship between rainfall data and criterion stays unchanged, the variations of SWI values are reduced significantly. This shows that there are possibilities to limit the range of SWI values of critical lines for sediment disasters on the basis of geological features.

Key words

early-warning, sediment disaster, rainfall threshold, soil water index, tank model